Andrew Bailey, Governor of the Bank of England, says rigorous testing should come before regulation as the starting point for managing AI risks. He argues that AI development should continue because its potential benefits are substantial, while safeguards must allow authorities to intervene and set boundaries.

Bailey says AI risks are real and growing, and that the pace of UK testing needs to accelerate. The UK’s AI Security Institute is contributing to work on standards; tests may expose unexpected behaviour, which he says demonstrates why testing is needed. Bailey leaves open the possibility of a formal regulatory framework later, while noting that AI firms in the US have signed a voluntary, “morally binding” safety agreement.